Protein Micro Arrays

7/31/2019 Protein Micro Arrays

1/10

REVIEW

Protein microarrays for diagnostic assays

Michael Hartmann & Johan Roeraade & Dieter Stoll &Markus F. Templin & Thomas O. Joos

Received: 24 June 2008 /Revised: 6 August 2008 /Accepted: 1 September 2008 /Published online: 20 September 2008# Springer-Verlag 2008

Abstract Protein microarray technology has enormous po-

tential for in vitro diagnostics (IVD). Miniaturized parallelizedimmunoassays are perfectly suited to generating a maximum

of diagnostically relevant information from minute amounts

of sample whilst only requiring small amounts of reagent.

Protein microarrays have become well-established research

tools in basic and applied research and the first products are

already on the market. This article reviews the current state of

protein microarrays and discusses developments and future

demands relating to protein arrays in their role as multiplexed

immunoassays in the field of diagnostics.

Keywords Protein microarrays . Multiplexed diagnostics .

In-vitro diagnostics . Focused protein-profiling .

Analytical microarrays

Introduction

Ambient analyte theory

Immunoassays have been widely employed as highly sensi-

tive tools for almost half a century [1, 2]. Antibody-based

immunoassays enable the generation of highly robust assays

that can be easily standardized and automated. Nowadays

there are hundreds of antibody-based assays on the diagnos-

tic market [3]. The miniaturization of immunoassays fordiagnostic purposes started way back in the early sixties [4,

5]. Feinberg and his colleagues developed a microspot-based

assay that enabled them to diagnose autoimmune diseases.

Auto-antigens were immobilized in a microspot and incu-

bated with human serum. The presence of auto-antibodies in

the serum led to the spontaneous precipitation of the auto-

antigens in the microspot. The authors assumed that such

microspot assays would have a particular advantage for

routine use on clinical specimens because it is simple,

sensitive, objective, quickly carried out and read, requires but

minute quantities of serum and antigen, and provides a

permanent record for the case files. However, this micro-

spot-based assay format did not have many supporters. More

than two decades later, Roger Ekins came up with the

ambient analyte theory on the feasibility of highly sensitive

multi-spot and multi-analyte immunoassays [6]. According

to Ekins theory, miniaturization also leads to an increase in

detection sensitivity1. According to the law of mass action,

only a small amount of analyte molecules will be captured on

to a spot that only contains a minute amount of immobilized

capture molecules. Under ambient analyte conditions, the

concentration of analyte in the sample will not change

significantly even though the capture molecule is a high-

affinity binder and the analyte concentration is low (Fig. 1).

Therefore, this type of miniature immunoassay is concentra-

tion-dependent: the analyte molecules captured in the spot

directly reflect the analyte concentration in the sample. In

consequence of the unaltered analyte concentration, the signal

becomes independent of the sample volume. High sensitivity

Anal Bioanal Chem (2009) 393:14071416

DOI 10.1007/s00216-008-2379-z

M. Hartmann : J. Roeraade

School of Chemical Science and Engineering,

Department of Analytical Chemistry,

Royal Institute of Technology,

10044 Stockholm, Sweden

M. Hartmann : D. Stoll : M. F. Templin : T. O. Joos (*)

NMI Natural and Medical Sciences

Institute at the University of Tbingen,

72770 Reutlingen, Germany

e-mail: [email protected]

1 Here detection sensitivity is used in terms of limit of detection, i.e.

the smallest amount of analyte which can be detected with acceptable

statistical significance. By contrast we use the term assay sensitivity,

i.e. the smallest detectable difference in analyte concentration.


2/10

is achieved because the measurement always takes place at the

highest analyte concentration possible. Thus, for a given

analyte concentration, it is possible that the detection

sensitivity in a microspot is higher than in a macrospot,

because the analyte is contained in a small area and only signaldensity (i.e. signal per area) not the total sum of signal is

relevant for signal-to-noise ratios defining the lower limit of

detection. For this reason and the possibility of being able to

detect multiple analytes in a single experiment, Ekins was

convinced that miniaturized and parallelized protein micro-

arrays had enormous potential for diagnostic applications [6].

From genomics to proteomics

Despite protein microarrays huge potential for diagnostic

applications, it was, nevertheless, the field of genomics that

made the greatest contribution to the advance of microarray

technology. Following the sequencing of the entire human

genome, DNA microarrays were developed and applied to

large-scale genomics research. Comprising up to tens of

thousands of different oligonucleotide probes per square

centimeter, DNA microarrays are perfectly suited to high-

throughput hybridization systems that enable the expression

analysis of the entire transcriptome in a single experiment

[7, 8]. Nowadays, DNA microarrays are well-established

and reliable methods for mRNA expression profiling and

SNP analysis. In addition, the first diagnostic assays have

been placed on the market [9, 10]. Despite the huge potential

of DNA, it is nevertheless proteins that are the key players in

cellular processes. DNA microarrays only provide a limited

amount of information on the status of cells and tissues.

Cellular functions depend on protein activity, and proteins

are not only regulated by the differential expression of the

underlying genes but also by post-translational modifica-tions. Furthermore, protein expression does not generally

quantitatively correlate with mRNA expression [11]. Within

the last decade, protein microarrays have entered the field

of proteomic research [12, 13] demonstrating their huge

capacity for identifying and quantifying proteins and for

studying the function of proteins from the perspective of

the proteome as a whole [14]. Besides their application

in proteome-wide research and the determination of the

biochemical activities of proteins, protein microarrays are

also perfect tools for quantitating subsets of proteins in com-

plex mixtures. This approach is known as focused protein-

profiling or analytical microarrays. As correctly foreseen byRoger Ekins, multiplex immunoassays like these have been

adapted to the microarray format and are about to be used in

diagnostic applications. This article reviews the present state

of multiplex immunoassays and discusses the obstacles that

still slightly hamper their breakthrough in IVD.

Protein microarrays

Principles and basics

DNA microarrays preceded the use of protein microarrays.

When proteins rather than DNA are used, a variety of

additional challenges arise due to the more complex nature

of proteins. A DNA molecule is built up from four different

bases that all have the same hydrophilic sugar backbone.

DNA is a very uniform and stable molecule due to its

chemical structure and the pairing of complementary bases.

DNA molecules exhibit a strong one-to-one interaction,

which, under physiological conditions, is biochemically

more or less inert. Proteins, which are a lot more fragile, are

assembled from 20 different amino acids that are extremely

heterogeneous in size and can have either hydrophobic or

hydrophilic side chains. The proper functioning of proteins

depends on their tertiary and quaternary structures. The

foundation of these structures is a balanced system of

electrostatic forces, hydrogen bonds, and hydrophobic Van

der Waals interactions. In contrast with DNA, slight changes

in salt concentration or pH, the presence of oxidants, or the

removal of water often irreversibly harms their structure and

hence interferes with their function. Furthermore, PCR and the

ability to chemically synthesize oligonucleotides made a con-

siderable contribution to the success of DNA microarray

analytics. The amplification of proteins is impossible. High-

Fig. 1 Signal and signal density in microspots. Signal density (signal/

area) and signal (total intensity) of captured targets in microspots are

shown for different concentrations of capture molecules. The capture

molecules are immobilized with the same surface density on all spots.

The signal (total signal) increases with increasing amount of capture

molecules for growing spot size. When most of the targets are capturedfrom the solution the signal reaches its maximum. By contrast, signal

density (signal/area) increases with decreasing amount of capture

molecules (decreasing spot size), reaching a constant level when the

capture molecule concentration is


3/10

affinity and high-specificity capture molecules can be easily

predicted and generated from the primary sequence of the

target DNA. Protein capture molecules cannot be predicted or

designed on the basis of a similarly easy principle, because the

interaction underlies a broad range of the aforementioned

molecular forces. In addition, post-translational modifications

such as phosphorylation, acetylation, or glycosylation have to

be taken into account when designing capture or bindingmolecules. At present, the lack of highly specific and high-

affinity capture molecules is still the main limitation of protein

microarrays.

Antibodies are broadly used in research and diagnostic

applications because of their ability to bind to target

proteins with high specificity and high affinity. When

protein microarrays were still in their infancy, scientists

initially resorted to antibodies since these proteins were

well known from immunoassays [1]. The invention of

monoclonal antibodies was a major step forward in the

generation of unlimited resources of defined capture

molecules. From then on, it was possible to produce pureand highly specific antibodies against almost any type of

antigen [15]. However, the generation and validation of

antibodies, regardless of whether they are monoclonal or

polyclonal, is a time-intensive and cost-intensive process.

In-vitro strategies that enabled the generation of binding

molecules were eventually developed. For instance, phage-

display technology can be used to screen large synthetic

libraries of protein clones, within a few weeks, for detection

of suitable binders against a target molecule of interest [16].

However, it also turned out that additional maturation steps

were needed for generation of high-affinity binders that

were similar to monoclonal or polyclonal antibodies.

Another promising strategy for producing synthetic binders

is the generation of aptamers. Aptamers are short single-

stranded nucleic acid oligomers (ssDNA or RNA) with a

specific and complex three-dimensional shape which causes

their well-fitting binding. Aptamers are produced using an

in-vitro selection and amplification technique called

SELEX (systematic evolution of ligands by exponential

enrichment). Many of the selected aptamers show affinities

comparable with those observed for monoclonal antibodies

[17]. Affibodies are another class of binding molecule.

They are based on combinatorial protein engineering of the

small and robust -helical structure of the domains of

protein A [18, 19]. These synthetic binders are more robust

than antibodies, which makes them perfectly suitable as

capture or detection agents in protein microarrays. However,

they had to prove that they had an affinity and specificity that

was similar to that of antibodies. Despite all the advances in

recombinant and scaffold-based technologies, the most

advanced binding reagent project today, the Human Protein

Atlas, uses a polyclonal antibody approach to generate

binding reagents against human proteins [20].

Planar microarrays

Protein microarrays are highly miniaturized and parallelized

solid-phase assay systems that use a large number of

different capture molecules immobilized in microspots

(diameter


4/10

based systems are similar to those obtained with established

ELISA systems involving planar arrays. Some bead-basedimmunoassays proved to be as sensitive, accurate, and

precise as competitive radio immunoassays [26]. Nowa-

days, the most popular platform is Luminexs xMAP

technology. xMAP differentiates 100 different color-coded

beads and allows researchers to easily set-up multiplex

assays with small numbers of analytes; alternatively, they

can choose from the increasing number of commercially

available kits (http://www.luminexcorp.com ). The BD

FACSArray Bioanalyzer (http://www.bdbiosciences.com )

or standard FACS instruments are able to classify the

different bead types by their size and can use two or more

different excitation lasers for multiplexed detection. This

allows the design of more complex assays. At present, the

washing steps in bead-based assays mainly involve filter

plates. However, the implementation of magnetic beads will

further simplify automation; problems experienced with

filter plates such as clogging, leaking, or unspecific

adsorption of analyte on to the large surface of filters will

be avoided. The integration of established magnetic bead-

handling technologies such as magnetic plate separators

[27], magnetic pin heads (www.thermo.com/kingfisher), or

in-tip magnetic capture (http://www.magbio.com) are a

decisive step towards the further automation of bead arrays.

Assay formats

Besides protein expression analysis, protein microarrays

can also be used for the functional analysis of proteins,

including protein interaction involving immobilized pro-teins or peptides, low molecular weight compounds, DNA,

oligosaccharides, tissues, or cells. In general, protein array

assay formats can be divided into forward-phase and

reversed-phase arrays (Fig. 3). Forward-phase protein

microarray assays involve the immobilization of capture

agents and hence enable the analysis of multiple parameters

from a single sample that is incubated on the array. In

reversed-phase protein microarrays, many different samples

(cell or tissue lysates) are immobilized in a microarray

format and are simultaneously analyzed for the presence of

a single target protein using a target-specific antibody.

Replicates of such protein arrays allow the analysis ofhundreds of parameters from minimal amounts of sample.

The reversed-phase array format is ideally suited to looking

into large sample cohorts. It enables the detection of

differentially regulated proteins in healthy or diseased

tissue and treated and untreated cells, the identification of

disease-specific biomarkers, or the analysis of cell signaling

networks [28]. Tissue arrays are a special type of reversed-

phase microarray and consist of tissue slices that are

immobilized on a surface. They are the miniature equiva-

lent of classical immunohistochemistry assays and enable

Fig. 2 Planar and bead-based microarrays. In planar microarrays,

individual capture agents are immobilized in a microarray format

containing between several hundreds to several thousands of spots.

The arrays are probed with sample and the analytes of interest bind to

their cognate capture agent. The binding reaction is verified by a

fluorescence read out. In bead-based microarrays, individual capture

agents are bound to color-coded or size-coded microspheres. The

assay can be performed with standard laboratory ware; a flow

cytometer is used to detect the fluorescent label

Fig. 3 Forward and reversed-phase array format. In a forward-phase

array format, a large number of immobilized capture molecules enable

the analysis of many parameters from a single sample, regardless

whether the sample is directly labeled or a sandwich assay is

performed. In a reversed-phase array, many samples are immobilized

and a highly specific antibody used to analyze the expression of a

single parameter in the immobilized samples

1410 M. Hartmann et al.
http://www.luminexcorp.com/http://www.bdbiosciences.com/http://www.thermo.com/kingfisherhttp://www.magbio.com/http://www.magbio.com/http://www.thermo.com/kingfisherhttp://www.bdbiosciences.com/http://www.luminexcorp.com/


5/10

the simultaneous analysis of large numbers of tissues with

minimum reagent consumption [29].

Currently, forward-phase assays are the most frequently

used protein microarray format. Protein arrays are used for

detailed expression analysis and for functional protein

studies. Forward-phase protein microarray assays are

mainly applied for antibody arrays to quantify dozens of

analytes that are present in complex samples. In antibodyarrays, antibodies are immobilized on a carrier at densities

of a few up to several thousand antibodies. Captured

analytes can either be visualized by directly labeling the

samples (analogous to DNA microarrays) or by multiplexed

sandwich immunoassays. Protein microarrays enable

screening for disease-related up or down-regulation of

proteins in patient samples [30]. Fluorescence labeling of

a sample necessitates careful optimization in order to obtain

optimum signal-to-noise ratios. Insufficient labeling proce-

dures result in a decrease in detection sensitivity whereas

over-labeling of a sample will cause high background

signals. The sample can also be labeled with biotin andsubsequently incubated with fluorescence-labeled streptavi-

din; this might improve signal-to-noise ratios [31]. How-

ever, it has to be kept in mind that the labeling of the target

proteins may interfere with the antibody-antigen interac-

tion, because the labeling procedure could negatively affect

the epitope and destroy it. Finally, one has to be aware that

proteins appear in complexes. A strong signal can thus

either be derived from large amounts of target analytes

captured or, alternatively, from capturing a huge protein

complex. The use of a single capture reagent cannot

discriminate between a specific and an unspecific binding

event. Specificity is greatly enhanced by using matched

antibody pairs in miniaturized sandwich immunoassays in

which high-affinity antibodies are immobilized in an array

format and capture the target analytes during incubation

with the sample. Unbound molecules are washed off

and the array is incubated with detection antibodies that

bind to another epitope on the target analytes. In sand-

wich immunoassays, two separate binding reactions are

employed; this is substantially more specific than direct

labeling approaches. The detection antibody can be directly

labeled with fluorescence; alternatively, the signal can be

further enhanced when using the biotin-streptavidin re-

porter system. The biotin-streptavidin model is a universal

reporter system that simplifies the assay procedure, because

all detection antibodies can be used in a biotinylated form.

Despite high selectivity and detection sensitivity, it is,

nevertheless, not possible to achieve unlimited multiplexing

of sandwich immunoassays due to cross reactivity arising

from the increasing overall concentration of the cocktail of

different detection antibodies, which increases background

signals above a certain threshold. In all complex multi-

plexed sandwich immunoassays discussed so far, multi-

plexed sandwich immunoassays used for the detection of

up to 100 analytes had to be divided into several multi-

plexed assays [32]. In an 11-plex sandwich immunoassay,

the detection sensitivity decreased by a factor of 1.7up to

5 compared with the single-plex assays, because higher

background signals wer e observ ed in the multip lex

format [33].

Another forward-phase assay type uses recombinantlyexpressed proteins that are immobilized as capture mole-

cules in a microspot. For example, in a study on protein

interaction, Zhu et al. arrayed 5,800 recombinant yeast

proteins on a microscope slide and probed this yeast

proteome microarray with proteins and phospholipids to

identify new interaction partners [14]. They confirmed the

already known existence of calmodulin-interacting and

phospholipid-interacting proteins and identified new inter-

acting proteins. These types of protein microarray can also

be used as antigen arrays to study the cross reactivity of

antibodies or to screen for autoantibodies against unknown

types of antigen. They represent a special forward-phaseassay type, because the immobilized antigens are not real

capture agents; instead the antibodies under analysis bind to

the immobilized antigens. Antigen microarrays are easy to

establish, because only one species-specific detection anti-

body is needed to identify bound antibodies in the microspots.

Antigen microarrays are ideally suited to the multiparametric

challenges of autoimmune and allergy diagnostics. Several

autoimmune and allergy microarray assays are already

commercially available and will be discussed in more detail

in the next section.

Other analytical tools

Apart from the above mentioned classical microarray

methods, other technologies are employed to analyze

protein interactions in a multiplex fashion. Very interesting

are label-free detection methods, because there is no need

for extra detection agents. For example, Quadraspecs

spinning disc interferometry (SDI) technology enables

detection of up to 128 unique analytes in 264 samples per

disk. They set up a veterinary diagnostic test for canine

heartworm (www.quadraspec.com). Analytik Jena com-

bined reflectometric interference spectroscopy (RIfS) with

biochip technology in their BIAffinity (www.analytik-jena.

de). Maven Biotechnologies (www.mavenbiotech.com)

LFIRE is an imaging system based on total internal

reflection ellipsometry that measures molecular binding

reactions in a microarray. According to the company,

LFIRE has been validated with protein microarray densities

of 2,500 spots cm2 and is capable of detecting molecules

as small as 150 Daltons. Finally, well established surface

plasmon resonance (SPR) technology has been adopted to

microarrays: in research applications, the Biacore Flexchip

Protein microarrays for diagnostic assays 1411
http://www.quadraspec.com/http://www.analytik-jena.de/http://www.analytik-jena.de/http://www.mavenbiotech.com/http://www.mavenbiotech.com/http://www.analytik-jena.de/http://www.analytik-jena.de/http://www.quadraspec.com/


6/10

has been used to detect up to 400 reactions in parallel [ 34,

35]. Apart from optical devices there are also electrochem-

ical platforms, for example the GRAVI-Chip from Diag-

noSwiss that uses electrochemical read out of enzymatic

reactions to detect a current that is proportionate to the

analyte concentration. Basically, many label-free technolo-

gies allow real-time detection of analytes and, therefore, the

analysis of kinetic parameters of analytes could giveadditional diagnostic information. However, label-free

technologies have to prove to be reliable enough to enter

the market of multiplexed applications.

This has already been proven by the Triage system (www.

biosite.com) and the VIDAS platform (www.biomerieux-

usa.com). Both are micro-fluidic and integrated devices

based on fluorescent read out that offer a set of multiplex

assays for detection of proteins related to cardiac diseases

(Table 1). The development of technology is also driven by

reducing costs for diagnostics. For example, common

recordable compact disks as molecular screening surfaces

and a standard optical CD/DVD drive as detector, havebeen reported [36, 37]. Alpha-fetoprotein has been detected

in a buffer environment by enzyme or gold nanoparticle-

labeled antibodies that were used as tracers, forming a

precipitate on the sensing disk surface. Such low-cost devices

could also contribute to a better acceptance of microarrays in

IvD.

Protein microarrays for diagnostics

Analytical protein microarrays for detection of antibodies

A very early example of a multiplex assay system is the

MASTpette test chamber which was developed for multi-

parametric allergy testing in the 1980s [38]. Allergen-coated

cellulose threads were bonded in the test chamber andincubated with patient serum. Bound IgEs were detected

with enzyme-labeled anti IgE antibody using a chemilumi-

nescent reaction. Today, IVD MASTpette test chamber-

based test systems use as little as 230 L for testing 20

allergen reactions in one assay (www.invernessmedical.de).

The multiplex detection of allergens was transferred to

protein microarrays, which has led to further miniaturiza-

tion and parallelization. One of the first protein microarrays

for detection of IgE used crude allergen extracts that were

immobilized in microspots on modified glass slides. This

approach involved rolling circle amplification in order to

achieve sufficient sensitivity [39]. Hiller et al. [40] used 94purified allergen molecules representing the most common

allergen sources to generate an allergen array to screen for

specific IgEs from minimal amounts of plasma samples.

Purified allergens enabled the scientists to detect bound IgEs

without needing to amplify them. Nowadays, protein micro-

arrays enable the detection of allergen-specific IgE reactivity

Table 1 Commercial multiplexed and protein-based diagnostics

Indication Target Vendor Platform FDA cleared

Allergies Antibodies VBC-Genomics (www.vbc-genomics.at) Planar array

Allergies, celiac disease Antibodies INOVA Diagnostics (www.inovadx.com) Luminex +

Allergies, common Antibodies ImmuneTech (www.immunetech.com) Luminex +

Allergies, indoor allergens Antibodies INDOOR Biotechnologies (www.inbio.com) Luminex

Autoimmune Antibodies BioArray Solutions (www.bioarrays.com) Bead array +

Autoimmune Antibodies Biomedical Diagnostics (www.bmd-net.com) Luminex +

Autoimmune Antibodies INOVA Diagnostics (www.inovadx.com) Luminex +

Autoimmune Antibodies Zeus Scientific (www.zeusscientific.com) Luminex +

Autoimmune Antibodies Bio-Rad Laboratories (www.bio-rad.com ) Luminex +

Autoimmune Antibodies Whatman (www.whatman.com) Planar array

Cancer Proteins RBM (www.rulesbasedmedicine.com) Luminex

Cardiac, heart failure Proteins Biomrieux (www.biomerieux-diagnostics.com) VIDAS +

Cardiac, myocardial infarction Proteins Biosite (www.biosite.com) Triage system

Cardiac, shortness of breath Proteins Biosite (www.biosite.com) Triage system

Infectious disease Antibodies Bio-Rad Laboratories (www.bio-rad.com) Luminex +

Infectious disease, Epstein-Barr Antibodies Zeus Scientific (www.zeusscientific.com) Luminex

Infectious disease, FSME, Borrelia Antibodies Multimetrix (www.multimetrix.com ) Luminex

Infectious disease, Herpes virus Antibodies Focus Diagnostics (www.focusdx.com) Luminex +

Multiple immunoassays Proteins Randox (www.randox.com ) Evidence +

Neurologic, Alzheimers disease Proteins Innogenetics (www.innogenetics.com) Luminex

Typing, HLA Antibodies Tepnel (www.tepnel.com) Luminex +

Genetic tests still dominate multiplexed diagnostics, but proteins are catching up. So far, predominantly antibodies have been detected in serum

from patients suffering from allergies, autoimmune, or infectious diseases. Technically, the Luminex platform is the prevailing technology for

multiplexed protein diagnostics (Source: Multiplexed Diagnostics 2008; www.SelectBiosciences.com)

http://www.biosite.com/http://www.biosite.com/http://www.biomerieux-usa.com/http://www.biomerieux-usa.com/http://www.invernessmedical.de/http://www.vbc-genomics.at/http://www.inovadx.com/http://www.immunetech.com/http://www.inbio.com/http://www.bioarrays.com/http://www.bmd-net.com/http://www.inovadx.com/http://www.zeusscientific.com/http://www.bio-rad.com/http://www.whatman.com/http://www.rulesbasedmedicine.com/http://www.biomerieux-diagnostics.com/http://www.biosite.com/http://www.biosite.com/http://www.bio-rad.com/http://www.zeusscientific.com/http://www.multimetrix.com/http://www.focusdx.com/http://www.randox.com/http://www.innogenetics.com/http://www.tepnel.com/http://www.selectbiosciences.com/http://www.selectbiosciences.com/http://www.tepnel.com/http://www.innogenetics.com/http://www.randox.com/http://www.focusdx.com/http://www.multimetrix.com/http://www.zeusscientific.com/http://www.bio-rad.com/http://www.biosite.com/http://www.biosite.com/http://www.biomerieux-diagnostics.com/http://www.rulesbasedmedicine.com/http://www.whatman.com/http://www.bio-rad.com/http://www.zeusscientific.com/http://www.inovadx.com/http://www.bmd-net.com/http://www.bioarrays.com/http://www.inbio.com/http://www.immunetech.com/http://www.inovadx.com/http://www.vbc-genomics.at/http://www.invernessmedical.de/http://www.biomerieux-usa.com/http://www.biomerieux-usa.com/http://www.biosite.com/http://www.biosite.com/


7/10

with the same sensitivity and specificity as the diagnostic

technologies that are currently used on a routine basis [41].

Over the last ten years, antigen arrays with several hundred,

and even up to several thousand, immobilized antigens have

been used for detection of specific autoantibodies involved in

autoimmune diseases. Joos et al. [42] used complex protein

microarrays to detect up to 18 different rheumatic disease-

specific autoantibodies in human sera, achieving sensitivitiesand specificities that were similar to established ELISA

methods. This concept was further developed by implement-

ing peptide antigens on a microarray, which allowed the

detailed characterization of patients autoimmune statuses

[43]. Protein microarrays are perfectly suited to such types of

analysis, especially when the quantity of sample material is

limited. Sharp et al. [44] suggested the application of

autoantibody profiling to improve diagnosis and prediction

of disease onset and severity. Protein microarrays are valu-

able diagnostic tools for disease monitoring and therapy,

because new clinical information can be gained from

comprehensive autoantibody profiling, especially whencombined with other clinical parameters, e.g. cytokine levels.

Arrays of viral and microbial antigens are the third type of

antigen array. Mezzasoma et al. [45] successfully demon-

strated the detection of antibodies directed against Toxo-

plasma gondii, rubella virus, cytomegalovirus, and herpes

simplex virus types 1 and 2 (ToRCH antigens) in serum

samples. In a more comprehensive study, Waterboer et al.

[46] used bead-based microarrays to analyze 756 sera for

the presence of antibodies against 27 antigens derived from

human papillomaviruses. Recombinant GST-tagged fusion

proteins were bound to glutathione beads and incubated

with patient sera. Bound antibodies were visualized using

anti-human IgG-specific detection. The authors were able to

correlate the protein array results with those from ELISA-

based methods. However, multiplexed serology also en-

abled them to effectively detect weak antibody responses.

Gray et al. [47] used protein arrays containing immobilized

malaria antigens to profile serum from malaria patients and

found a correlation of the results with resistance to malaria.

The authors analyzed combinations of reactivity to different

antigens instead of individual reactivity and were able to

group the patients according to their increased development

of clinical immunity using hierarchical clustering. These

studies demonstrated the potential of antigen microarrays,

which can compete with ELISA tests in respect of sensitivity

and robustness. Antigen microarrays are perfect tools for

these types of application and support health personnel in

diagnosis and prognosis.

Antibody and reversed-phase microarrays

Similarly to DNA microarrays used for mRNA expression

analysis, protein-capture arrays are used in dual labeling

approaches to investigate relative protein abundance in two

differentially labeled samples. Haab et al. [30] demonstrat-

ed the potential of such protein capture arrays by analyzing

115 characterized antibody-antigen interactions. Capture

antibodies were immobilized on poly-L-lysine-coated glass

slides. Patient samples were labeled with Cy5 or Cy3 dye.

Defined mixtures containing Cy5-labeled antigens at

different concentrations were incubated with Cy3-labeledcontrol antigens. It was possible to achieve a correct linear

relationship between antigen concentration and assay signal

in about 20% of all interactions observed. Another 30% of

the interactions reflected the antigen concentrations. Direct

labeling approaches have already been commercialized by

several companies (www.biochipnet.com). It must however

be noted that sandwich immunoassays achieve higher spec-

ificities and usually higher detection and assay sensitivities

than direct labeling approaches.

Focused protein microarrays such as miniaturized sand-

wich immunoassays, either planar or bead-based, have

evolved into tools that deliver many clinical data of highdiagnostic and prognostic value. For example, lyzed breast

tumor biopsies samples were analyzed along with normal

tissue for 14 relevant marker proteins from only 50 g

protein using the bead-based Luminex xMAP system [48].

The expression profile obtained for estrogen receptor and

Her-2 (human epidermal growth factor receptor 2) exactly

matched the results of the immunohistochemistry tests that

are normally used for this type of application. Recently, the

expression of 11 soluble receptors was analyzed in patient

samples from 36 critically ill intensive care unit patients.

Hierarchical clustering analysis allowed the scientists to

group the patients into a sepsis and a trauma group [ 33].

The field of inflammation holds huge potential for protein

microarrays as diagnostic tools, especially in the field of

sepsis, which is often hampered by the lack of quantitative

IVD tests. Due to the complex nature of sepsis, many clinical

parameters ought to be monitored in parallel. Protein micro-

arrays are perfectly suited to fulfill this requirement [49].

Reversed-phase microarrays also have enormous poten-

tial as diagnostic tools, in particular in the identification of

disease-specific biomarkers, which may be used as markers

to initiate and monitor therapy. There are high expectations

that biopsy analysis from individual patients may lead to

therapies that are specifically tailored to individual require-

ments [50].

Commercial protein microarrays

A variety of analytical protein microarrays have been

developed for different platforms and are entering the

diagnostic market (Table 1). The first planar and bead-

based multiplex immunoassays have been cleared by the

US FDA or have been CE-marked for use in the EU. These

http://www.biochipnet.com/http://www.biochipnet.com/


8/10

protein arrays are produced by the manufacturers whilst the

assay and data analysis is performed by the customers. The

AtheNA Multi-Lyte test system (Zeus Scientific, Raritan,

USA) [51] and the BioPlex 2200 ANA screen (Bio-Rad,

Hercules, USA) [52] are designed for autoimmune testing

and are based on Luminexs xMAP technology. These

multiplexed immunoassays can be used to screen samples

for the presence of multiple autoantibodies involved inrheumatic diseases such as systemic lupus erythematosus

(SLE), mixed connective tissue disease, Sjgrens syn-

drome, scleroderma (systemic sclerosis), polymyositis, and

CREST syndrome. The AtheNA Multi-Lyte test system

uses intra-well calibration technology involving internal

standards for analyzing patient serum characteristics. The

measured assay signals are corrected to compensate assay

drift and to allow patient-specific calibration. A positive

result is achieved for each individual analyte test when the

assay signal is above a specific upper threshold. The result

is defined as negative when the assay signal is below a

defined threshold. Signals between the lower and upperthresholds are marked as questionable for this specific

analyte. The BioPlex 2200 ANA Screen system uses

pattern-recognizing medical decision support software that

associates test results with predefined patterns that have

been correlated with autoimmune diseases. The k-nearest

algorithm is used to check for greatest concordance with 11

reference samples from a database containing data sets

generated from more than 1,400 patient samples.

CombiChip Autoimmune 1.0 from Whatman (Spring-

field Mill, UK) is a planar microarray assay for autoim-

mune diagnostics. This protein microarray uses 14 different

autoantigens that are immobilized on nitrocellulose-coated

slides within 16 identical subarrays. Using this autoimmune

microarray, 16 patient samples per slide can be processed

using multi-channel pipettes in the same way as conventional

ELISA testing using microtiter plates. However, imaging and

image analysis has to be done manually using common micro-

array scanners and image analysis software. The CombiChip

Autoimmune is CE-marked and sold in the EU; in the USA,

the software is available for research use only.

Randoxs Laboratories (Crumlin, UK) have developed

Evidence, an automated biochip system enabling the

analysis of miniaturized and parallelized immunoassays in

a macroarray format containing 25 features and using

chemiluminescence-based read out [53]. Several multiplexed

immunoassays panels have been developed, including

fertility, cardiac disease, tumors, cytokines and growth

factors, cell adhesion molecules, thyroid function, and drug

residues panels. The drugs abuse array has already received

FDA clearance and other assays are currently being evaluated,

thus making the Evidence biochip analyzer a pioneer in

multiplexed immunoassays for clinical diagnostics.

Problems and requirements

Protein microarrays are well-established analytical tools in

basic and applied research, which is a sign of the

capabilities and power of these assay systems. Prior to

their implementation in routine clinical diagnostics, protein

microarray-based results have to demonstrate clinical

relevance in the initiation or changing of therapy. Medicaldemand, combined with an overall cost reduction, must

become the driving force behind protein arrays gaining a

substantial share in the IVDs market. Although protein

microarrays have huge diagnostic potential, they are

nevertheless still far from being widely used in IVDs.

Several regulatory hurdles have to be cleared, for example,

the IVD tests have to fulfill the specifications of the 98/97/

EC directive of the European Parliament and of the

European Council of 22/12/98. In the USA, the Food and

Drug Administration (FDA) decides on the approval of

IVDs for human application. Before approval is granted,

the tests have to provide valid results, and proof that theresults have a positive therapeutic outcome. In order to

evaluate the reliability and reproducibility of DNA micro-

arrays for gene expression analysis, the FDAs National

Center for Toxicological Research (NCTR) set up the

MicroArray Quality-Control (MAQC) project that involves

both academic and commercial partners. (http://www.fda.

gov/nctr/science/centers/toxicoinformatics/maqc ). In the

first phase, more than 1,300 microarray experiments were

performed on different platforms at different test sites with

the aim of showing that a consensus in data analysis will

make a considerable contribution to reproducible lists of

differentially expressed genes [54]. In the next phase, the

tests are further validated in terms of clinical applicability.

Special focus is put on data analysis using different

algorithms in order to obtain predictive signatures and

classifiers. However, DNA analysis is unable to provide

data of clinical relevance in the same way as protein

analysis. Only proteins, and hence the proteome, are able to

reflect the physiological state of a cell. DNA analysis

enables the diagnosis of a persons genetic predisposition to

a specific disease, but it will be not possible to predict the

onset of disease. Although analytical protein microarrays

are able to analyze fairly focused sets of analytes, a variety

of problems have to be solved prior to the broad application

of multiplexed protein microarrays on the IVD market. To

ensure the high quality of microarrays, much more effort

has to be made than with singleplex measurements. A

variety of controls have already been implemented in

protein microarrays, such as replicate spots, marker spots

for orientation, negative control spots to check for

unspecific binding, application of assay buffer to test for

cross reactivity between capture molecules and detection

http://www.fda.gov/nctr/science/centers/toxicoinformatics/maqchttp://www.fda.gov/nctr/science/centers/toxicoinformatics/maqchttp://www.fda.gov/nctr/science/centers/toxicoinformatics/maqchttp://www.fda.gov/nctr/science/centers/toxicoinformatics/maqc


9/10

antibodies, internal spot normalization [55, 56] and differ-

ent normalization strategies [57]. However, quality control

in terms of IVD testing means the routine analysis of

reference sample material of low, medium and high

concentration [58]. Only this will ensure that the test is

effective at any analyte concentration and is able to detect

an assay drift. At present, no information exists on the

required complexity of such a reference sample and howstability can be guaranteed. The generation of defined

mixtures of control analytes and the solubility of these

molecules at high concentrations might cause problems that

have not so far broadly investigated. Putting such controls

in place will make it easier for the assay to be used for

clinical analysis. However, additional issues will have to be

addressed, for example the interpretation of assay results in

cases when some of the controls fail. Can we regard the

partial results gained from microarray with valid controls as

sufficiently reliable or do these results have to be rejected,

also? Another important issue is multiple subarrays within

one slide. What happens if one of these subarrays fails?How can one tell whether the whole array is affected? How

should failures within replicate spots be dealt with?

Replicate spots are a compromise arising from the limited

space within an array. Roches new microarray platform,

Impact, employs up to 20 replicate spots [59] and bead-

based systems analyze from fifty to one hundred beads

[60]. In the case of a multiplexed diagnostic assay for the

analysis of dozens of individual parameters being estab-

lished and a few or even just one single test being changed,

will a whole new approval for IVD be required?

In addition, multiplexed assays generate huge data sets,

which require appropriate data analysis. The question will be

how to draw conclusions from patterns rather than looking

into single interactions. The MicroArray Quality Control

project (see above) was introduced to provide the necessary

information for inter-platform and inter-laboratory concor-

dance of DNA microarray experiments [54]. The BioPlex

2200 ANA screen system uses a pattern-recognition algo-

rithm for analysis of their multiplexed protein assays [52].

However, there will be no general approach on how to apply

such algorithms to generate diagnostically relevant data sets.

Besides the regulatory aspects, social and ethical issues

also need to be considered. Should a customer be allowed

to order only part of the data generated with a multiplexed

assay and therefore pay less than the price of the full

multiplexed panel? Can the manufacturer or the performer

sell only parts of the microarray results, for example by

using different software settings? What happens to unre-

quested data sets? What happens if the unrequested data

sets are of diagnostic relevance that would lead to a

different diagnosis? All these questions have to be carefully

addressed and solved by the regulatory offices and by the

diagnostic companies before protein microarrays are placed

on the IVD market.

Protein microarrays assays will have to be automated

before entering the IVD market. Automation increases assay

performance, robustness, and reliability of multiplexed

assays. However, such automated multiplex platforms have

to compete with the well-established clinical analyzers that

currently dominate the diagnostic market. These systems caneasily increase throughput, e.g. measuring five parameters

from the same sample in a sequential mode. Therefore, as

long as sample material is not limited, or multiplexing does

not exceed more than five parameters, the diagnostic

companies are hesitant to make huge investments in order

to change the assay format. In the field of autoimmune

disease diagnostics, multiplexed assays are currently enter-

ing the diagnostic market, and sets of tumor marker panels

may in future also be applied to monitor therapy. It can be

safely assumed that protein microarrays will find their place

in the IVD market in areas in which sample volume is limited

and where they deliver therapeutically relevant data sets.

References

1. Ekins RP (1998) Clin Chem 44:20152030

2. Yalow RS, Berson SA (1960) J Clin Invest 39:11571175

3. Saleem M (2008) Afr J Biotechnol 7:923925

4. Feinberg JG (1961) Nature 192:985986

5. Feinberg JG, Wheeler AW (1963) J Clin Pathol 16:282284

6. Ekins RP (1989) J Pharm Biomed Anal 7:155168

7. Hoheisel JD (2006) Nat Rev Genet 7:200210

8. Plomin R, Schalkwyk LC (2007) Dev Sci 10:1923

9. Ng JK, Liu WT (2006) Anal Bioanal Chem 386:427434

10. Wang L, Luhm R, Lei M (2007) Adv Exp Med Biol 593:105116

11. Gygi SP, Rochon Y, Franza BR, Aebersold R (1999) Mol Cell

Biol 19:17201730

12. MacBeath G (2002) Nat Genet 32:526532 Suppl

13. Templin MF, Stoll D, Schwenk JM, Potz O, Kramer S, Joos TO

(2003) Proteomics 3:21552166

14. Zhu H, Bilgin M, Bangham R, Hall D, Casamayor A, Bertone P,

Lan N, Jansen R, Bidlingmaier S, Houfek T, Mitchell T, Miller P,

Dean RA, Gerstein M, Snyder M (2001) Science 293:21012105

15. Kohler G, Milstein C (1975) Nature 256:495497

16. Knappik A, Ge L, Honegger A, Pack P, Fischer M, Wellnhofer G,

Hoess A, Wolle J, Pluckthun A, Virnekas B (2000) J Mol Biol

296:5786

17. Stoltenburg R, Reinemann C, Strehlitz B (2007) Biomol Eng

24:381403

18. Gunneriusson E, Nord K, Uhlen M, Nygren P (1999) Protein Eng

12:873878

19. Gunneriusson E, Samuelson P, Ringdahl J, Gronlund H, Nygren

PA, Stahl S (1999) Appl Environ Microbiol 65:41344140

20. Hober S, Uhlen M (2008) Curr Opin Biotechnol 19:3035

21. Zhu H, Snyder M (2003) Curr Opin Chem Biol 7:55 63

22. Angenendt P, Glokler J, Sobek J, Lehrach H, Cahill DJ (2003) J

Chromatogr A 1009:97104

23. Seurynck-Servoss SL, White AM, Baird CL, Rodland KD, Zangar

RC (2007) Anal Biochem 371:105115



10/10

24. Barbulovic-Nad I, Lucente M, Sun Y, Zhang M, Wheeler AR,

Bussmann M (2006) Crit Rev Biotechnol 26:237259

25. Templin MF, Stoll D, Bachmann J, Joos TO (2004) Comb Chem

High Throughput Screen 7:223229

26. Opalka D, Lachman CE, MacMullen SA, Jansen KU, Smith JF,

Chirmule N, Esser MT (2003) Clin Diagn Lab Immunol 10:108

115

27. Yu H (1998) Anal Chim Acta 376:7781

28. Hanash S (2003) Nature 422:226232

29. Kononen J, Bubendorf L, Kallioniemi A, Barlund M, Schraml P,Leighton S, Torhorst J, Mihatsch MJ, Sauter G, Kallioniemi OP

(1998) Nat Med 4:844847

30. Haab BB, Dunham MJ, Brown PO (2001) Genome Biol 2:

RESEARCH0004

31. Kusnezow W, Banzon V, Schroder C, Schaal R, Hoheisel JD,

Ruffer S, Luft P, Duschl A, Syagailo YV (2007) Proteomics

7:17861799

32. Schweitzer B, Roberts S, Grimwade B, Shao W, Wang M, Fu Q,

Shu Q, Laroche I, Zhou Z, Tchernev VT, Christiansen J, Velleca

M, Kingsmore SF (2002) Nat Biotechnol 20:359365

33. Hsu HY, Wittemann S, Schneider EM, Weiss M, Joos TO (2008)

Med Eng Phys doi:(10.1016/j.medengphy.2008.01.003) or PMID:

18313970 (http://www.ncbi.nlm.nih.gov/pubmed/18313970)

34. Baggio R, Carven GJ, Chiulli A, Palmer M, Stern LJ, Arenas JE

(2005) J Biol Chem 280:4188419435. Usui-Aoki K, Shimada K, Nagano M, Kawai M, Koga H (2005)

Proteomics 5:23962401

36. Morais S, Carrascosa J, Mira D, Puchades R, Maquieira A (2007)

Anal Chem 79:76287635

37. Morais S, Tamarit-Lopez J, Carrascosa J, Puchades R, Maquieira

A (2008) Anal Bioanal Chem 391:28372844

38. Brown CR, Higgins KW, Frazer K, Schoelz LK, Dyminski JW,

Marinkovich VA, Miller SP, Burd JF (1985) Clin Chem 31:1500

1505

39. Wiltshire S, OMalley S, Lambert J, Kukanskis K, Edgar D,

Kingsmore SF, Schweitzer B (2000) Clin Chem 46:19901993

40. Hiller R, Laffer S, Harwanegg C, Huber M, Schmidt WM,

Twardosz A, Barletta B, Becker WM, Blaser K, Breiteneder H,

Chapman M, Crameri R, Duchene M, Ferreira F, Fiebig H,

Hoffmann-Sommergruber K, King TP, Kleber-Janke T, Kurup VP,

Lehrer SB, Lidholm J, Muller U, Pini C, Reese G, Scheiner O,

Scheynius A, Shen HD, Spitzauer S, Suck R, Swoboda I, Thomas

W, Tinghino R, Van Hage-Hamsten M, Virtanen T, Kraft D,

Muller MW, Valenta R (2002) FASEB J 16:414416

41. Harwanegg C, Hiller R (2005) Clin Chem Lab Med 43:13211326

42. Joos TO, Schrenk M, Hopfl P, Kroger K, Chowdhury U, Stoll D,

Schorner D, Durr M, Herick K, Rupp S, Sohn K, Hammerle H

(2000) Electrophoresis 21:26412650

43. Robinson WH, DiGennaro C, Hueber W, Haab BB, Kamachi M,

Dean EJ, Fournel S, Fong D, Genovese MC, de Vegvar HE,

Skriner K, Hirschberg DL, Morris RI, Muller S, Pruijn GJ, van

Venrooij WJ, Smolen JS, Brown PO, Steinman L, Utz PJ (2002)

Nat Med 8:295301

44. Sharp V, Utz PJ (2007) Nat Clin Pract Rheumatol 3:96103

45. Mezzasoma L, Bacarese-Hamilton T, Di Cristina M, Rossi R,Bistoni F, Crisanti A (2002) Clin Chem 48:121130

46. Waterboer T, Sehr P, Michael KM, Franceschi S, Nieland JD, Joos

TO, Templin MF, Pawlita M (2005) Clin Chem 51:18451853

47. Gray JC, Corran PH, Mangia E, Gaunt MW, Li Q, Tetteh KK,

Polley SD, Conway DJ, Holder AA, Bacarese-Hamilton T, Riley

EM, Crisanti A (2007) Clin Chem 53:12441253

48. Schneiderhan-Marra N, Kirn A, Dttinger A, Templin M, Sauer

G, Deissler H, Joos TO (2005) CGP 2:3742

49. Heuer JG, Cummins DJ, Edmonds BT (2005) Expert Rev

Proteomics 2:669680

50. Speer R, Wulfkuhle J, Espina V, Aurajo R, Edmiston KH, Liotta

LA, Petricoin EF (2007) Cancer Genomics Proteomics 4:157164

51. Gilburd B, Abu-Shakra M, Shoenfeld Y, Giordano A, Bocci EB,

delle Monache F, Gerli R (2004) Clin Dev Immunol 11:5356

52. Shovman O, Gilburd B, Barzilai O, Shinar E, Larida B, Zandman-Goddard G, Binder SR, Shoenfeld Y (2005) Ann N Y Acad Sci

1050:380388

53. Fitzgerald SP, Lamont JV, McConnell RI, Benchikh el O (2005)

Clin Chem 51:11651176

54. Casciano DA, Woodcock J (2006) Nat Biotechnol 24:1103

55. Hartmann M, Schrenk M, Dottinger A, Nagel S, Roeraade J, Joos

TO, Templin MF (2008) Clin Chem 54:956963

56. Olle EW, Sreekumar A, Warner RL, McClintock SD, Chinnaiyan

AM, Bleavins MR, Anderson TD, Johnson KJ (2005) Mol Cell

Proteomics 4:16641672

57. Hamelinck D, Zhou H, Li L, Verweij C, Dillon D, Feng Z, Costa

J, Haab BB (2005) Mol Cell Proteomics 4:773784

58. Master SR, Bierl C, Kricka LJ (2006) Drug Discov Today

11:10071011

59. Hornauer H, Klause U, Mller H, Vieth F, Risse B (2004)

Biospektrum 10:564565

60. Fulton RJ, McDade RL, Smith PL, Kienker LJ, Kettman JR

(1997) Clin Chem 43:17491756

61. Templin MF, Stoll D, Schrenk M, Traub PC, Vohringer CF, Joos

TO (2002) Trends Biotechnol 20:160166

http://dx.doi.org/10.1016/j.medengphy.2008.01.003http://www.ncbi.nlm.nih.gov/pubmed/18313970http://www.ncbi.nlm.nih.gov/pubmed/18313970http://dx.doi.org/10.1016/j.medengphy.2008.01.003

Documents

Protein Micro Arrays